man_validate.c

   1 /*      $OpenBSD$ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2010, 2012, 2013, 2014 Ingo Schwarze <schwarze@openbsd.org>
   5  *
   6  * Permission to use, copy, modify, and distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18 #include "config.h"
  19
  20 #include <sys/types.h>
  21
  22 #include <assert.h>
  23 #include <ctype.h>
  24 #include <errno.h>
  25 #include <limits.h>
  26 #include <stdarg.h>
  27 #include <stdlib.h>
  28 #include <string.h>
  29 #include <time.h>
  30
  31 #include "man.h"
  32 #include "mandoc.h"
  33 #include "mandoc_aux.h"
  34 #include "libman.h"
  35 #include "libmandoc.h"
  36
  37 #define CHKARGS   struct man *man, struct man_node *n
  38
  39 typedef void    (*v_check)(CHKARGS);
  40
  41 static  void      check_eq0(CHKARGS);
  42 static  void      check_eq2(CHKARGS);
  43 static  void      check_le1(CHKARGS);
  44 static  void      check_le5(CHKARGS);
  45 static  void      check_par(CHKARGS);
  46 static  void      check_part(CHKARGS);
  47 static  void      check_root(CHKARGS);
  48 static  void      check_text(CHKARGS);
  49
  50 static  void      post_AT(CHKARGS);
  51 static  void      post_IP(CHKARGS);
  52 static  void      post_vs(CHKARGS);
  53 static  void      post_fi(CHKARGS);
  54 static  void      post_ft(CHKARGS);
  55 static  void      post_nf(CHKARGS);
  56 static  void      post_TH(CHKARGS);
  57 static  void      post_UC(CHKARGS);
  58 static  void      post_UR(CHKARGS);
  59
  60 static  v_check man_valids[MAN_MAX] = {
  61         post_vs,    /* br */
  62         post_TH,    /* TH */
  63         NULL,       /* SH */
  64         NULL,       /* SS */
  65         NULL,       /* TP */
  66         check_par,  /* LP */
  67         check_par,  /* PP */
  68         check_par,  /* P */
  69         post_IP,    /* IP */
  70         NULL,       /* HP */
  71         NULL,       /* SM */
  72         NULL,       /* SB */
  73         NULL,       /* BI */
  74         NULL,       /* IB */
  75         NULL,       /* BR */
  76         NULL,       /* RB */
  77         NULL,       /* R */
  78         NULL,       /* B */
  79         NULL,       /* I */
  80         NULL,       /* IR */
  81         NULL,       /* RI */
  82         post_vs,    /* sp */
  83         post_nf,    /* nf */
  84         post_fi,    /* fi */
  85         NULL,       /* RE */
  86         check_part, /* RS */
  87         NULL,       /* DT */
  88         post_UC,    /* UC */
  89         check_le1,  /* PD */
  90         post_AT,    /* AT */
  91         NULL,       /* in */
  92         post_ft,    /* ft */
  93         check_eq2,  /* OP */
  94         post_nf,    /* EX */
  95         post_fi,    /* EE */
  96         post_UR,    /* UR */
  97         NULL,       /* UE */
  98         NULL,       /* ll */
  99 };
 100
 101
 102 void
 103 man_valid_post(struct man *man)
 104 {
 105         struct man_node *n;
 106         v_check         *cp;
 107
 108         n = man->last;
 109         if (n->flags & MAN_VALID)
 110                 return;
 111         n->flags |= MAN_VALID;
 112
 113         switch (n->type) {
 114         case MAN_TEXT:
 115                 check_text(man, n);
 116                 break;
 117         case MAN_ROOT:
 118                 check_root(man, n);
 119                 break;
 120         case MAN_EQN:
 121                 /* FALLTHROUGH */
 122         case MAN_TBL:
 123                 break;
 124         default:
 125                 cp = man_valids + n->tok;
 126                 if (*cp)
 127                         (*cp)(man, n);
 128                 break;
 129         }
 130 }
 131
 132 static void
 133 check_root(CHKARGS)
 134 {
 135
 136         assert((man->flags & (MAN_BLINE | MAN_ELINE)) == 0);
 137
 138         if (NULL == man->first->child)
 139                 mandoc_msg(MANDOCERR_DOC_EMPTY, man->parse,
 140                     n->line, n->pos, NULL);
 141         else
 142                 man->meta.hasbody = 1;
 143
 144         if (NULL == man->meta.title) {
 145                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 146                     n->line, n->pos, NULL);
 147
 148                 /*
 149                  * If a title hasn't been set, do so now (by
 150                  * implication, date and section also aren't set).
 151                  */
 152
 153                 man->meta.title = mandoc_strdup("");
 154                 man->meta.msec = mandoc_strdup("");
 155                 man->meta.date = man->quick ? mandoc_strdup("") :
 156                     mandoc_normdate(man->parse, NULL, n->line, n->pos);
 157         }
 158 }
 159
 160 static void
 161 check_text(CHKARGS)
 162 {
 163         char            *cp, *p;
 164
 165         if (MAN_LITERAL & man->flags)
 166                 return;
 167
 168         cp = n->string;
 169         for (p = cp; NULL != (p = strchr(p, '\t')); p++)
 170                 mandoc_msg(MANDOCERR_FI_TAB, man->parse,
 171                     n->line, n->pos + (p - cp), NULL);
 172 }
 173
 174 #define INEQ_DEFINE(x, ineq, name) \
 175 static void \
 176 check_##name(CHKARGS) \
 177 { \
 178         if (n->nchild ineq (x)) \
 179                 return; \
 180         mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line, n->pos, \
 181             "line arguments %s %d (have %d)", \
 182             #ineq, (x), n->nchild); \
 183 }
 184
 185 INEQ_DEFINE(0, ==, eq0)
 186 INEQ_DEFINE(2, ==, eq2)
 187 INEQ_DEFINE(1, <=, le1)
 188 INEQ_DEFINE(5, <=, le5)
 189
 190 static void
 191 post_UR(CHKARGS)
 192 {
 193
 194         if (MAN_HEAD == n->type && 1 != n->nchild)
 195                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 196                     n->pos, "line arguments eq 1 (have %d)", n->nchild);
 197         check_part(man, n);
 198 }
 199
 200 static void
 201 post_ft(CHKARGS)
 202 {
 203         char    *cp;
 204         int      ok;
 205
 206         if (0 == n->nchild)
 207                 return;
 208
 209         ok = 0;
 210         cp = n->child->string;
 211         switch (*cp) {
 212         case '1':
 213                 /* FALLTHROUGH */
 214         case '2':
 215                 /* FALLTHROUGH */
 216         case '3':
 217                 /* FALLTHROUGH */
 218         case '4':
 219                 /* FALLTHROUGH */
 220         case 'I':
 221                 /* FALLTHROUGH */
 222         case 'P':
 223                 /* FALLTHROUGH */
 224         case 'R':
 225                 if ('\0' == cp[1])
 226                         ok = 1;
 227                 break;
 228         case 'B':
 229                 if ('\0' == cp[1] || ('I' == cp[1] && '\0' == cp[2]))
 230                         ok = 1;
 231                 break;
 232         case 'C':
 233                 if ('W' == cp[1] && '\0' == cp[2])
 234                         ok = 1;
 235                 break;
 236         default:
 237                 break;
 238         }
 239
 240         if (0 == ok) {
 241                 mandoc_vmsg(MANDOCERR_FT_BAD, man->parse,
 242                     n->line, n->pos, "ft %s", cp);
 243                 *cp = '\0';
 244         }
 245
 246         if (1 < n->nchild)
 247                 mandoc_vmsg(MANDOCERR_ARGCOUNT, man->parse, n->line,
 248                     n->pos, "want one child (have %d)", n->nchild);
 249 }
 250
 251 static void
 252 check_part(CHKARGS)
 253 {
 254
 255         if (MAN_BODY == n->type && 0 == n->nchild)
 256                 mandoc_msg(MANDOCERR_ARGCWARN, man->parse, n->line,
 257                     n->pos, "want children (have none)");
 258 }
 259
 260 static void
 261 check_par(CHKARGS)
 262 {
 263
 264         switch (n->type) {
 265         case MAN_BLOCK:
 266                 if (0 == n->body->nchild)
 267                         man_node_delete(man, n);
 268                 break;
 269         case MAN_BODY:
 270                 if (0 == n->nchild)
 271                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 272                             man->parse, n->line, n->pos,
 273                             "%s empty", man_macronames[n->tok]);
 274                 break;
 275         case MAN_HEAD:
 276                 if (n->nchild)
 277                         mandoc_vmsg(MANDOCERR_ARG_SKIP,
 278                             man->parse, n->line, n->pos,
 279                             "%s %s%s", man_macronames[n->tok],
 280                             n->child->string,
 281                             n->nchild > 1 ? " ..." : "");
 282                 break;
 283         default:
 284                 break;
 285         }
 286 }
 287
 288 static void
 289 post_IP(CHKARGS)
 290 {
 291
 292         switch (n->type) {
 293         case MAN_BLOCK:
 294                 if (0 == n->head->nchild && 0 == n->body->nchild)
 295                         man_node_delete(man, n);
 296                 break;
 297         case MAN_BODY:
 298                 if (0 == n->parent->head->nchild && 0 == n->nchild)
 299                         mandoc_vmsg(MANDOCERR_PAR_SKIP,
 300                             man->parse, n->line, n->pos,
 301                             "%s empty", man_macronames[n->tok]);
 302                 break;
 303         default:
 304                 break;
 305         }
 306 }
 307
 308 static void
 309 post_TH(CHKARGS)
 310 {
 311         struct man_node *nb;
 312         const char      *p;
 313
 314         check_le5(man, n);
 315
 316         free(man->meta.title);
 317         free(man->meta.vol);
 318         free(man->meta.source);
 319         free(man->meta.msec);
 320         free(man->meta.date);
 321
 322         man->meta.title = man->meta.vol = man->meta.date =
 323             man->meta.msec = man->meta.source = NULL;
 324
 325         nb = n;
 326
 327         /* ->TITLE<- MSEC DATE SOURCE VOL */
 328
 329         n = n->child;
 330         if (n && n->string) {
 331                 for (p = n->string; '\0' != *p; p++) {
 332                         /* Only warn about this once... */
 333                         if (isalpha((unsigned char)*p) &&
 334                             ! isupper((unsigned char)*p)) {
 335                                 mandoc_vmsg(MANDOCERR_TITLE_CASE,
 336                                     man->parse, n->line,
 337                                     n->pos + (p - n->string),
 338                                     "TH %s", n->string);
 339                                 break;
 340                         }
 341                 }
 342                 man->meta.title = mandoc_strdup(n->string);
 343         } else {
 344                 man->meta.title = mandoc_strdup("");
 345                 mandoc_msg(MANDOCERR_TH_NOTITLE, man->parse,
 346                     nb->line, nb->pos, "TH");
 347         }
 348
 349         /* TITLE ->MSEC<- DATE SOURCE VOL */
 350
 351         if (n)
 352                 n = n->next;
 353         if (n && n->string)
 354                 man->meta.msec = mandoc_strdup(n->string);
 355         else {
 356                 man->meta.msec = mandoc_strdup("");
 357                 mandoc_vmsg(MANDOCERR_MSEC_MISSING, man->parse,
 358                     nb->line, nb->pos, "TH %s", man->meta.title);
 359         }
 360
 361         /* TITLE MSEC ->DATE<- SOURCE VOL */
 362
 363         if (n)
 364                 n = n->next;
 365         if (n && n->string && '\0' != n->string[0]) {
 366                 man->meta.date = man->quick ?
 367                     mandoc_strdup(n->string) :
 368                     mandoc_normdate(man->parse, n->string,
 369                         n->line, n->pos);
 370         } else {
 371                 man->meta.date = mandoc_strdup("");
 372                 mandoc_msg(MANDOCERR_DATE_MISSING, man->parse,
 373                     n ? n->line : nb->line,
 374                     n ? n->pos : nb->pos, "TH");
 375         }
 376
 377         /* TITLE MSEC DATE ->SOURCE<- VOL */
 378
 379         if (n && (n = n->next))
 380                 man->meta.source = mandoc_strdup(n->string);
 381         else if (man->defos != NULL)
 382                 man->meta.source = mandoc_strdup(man->defos);
 383
 384         /* TITLE MSEC DATE SOURCE ->VOL<- */
 385         /* If missing, use the default VOL name for MSEC. */
 386
 387         if (n && (n = n->next))
 388                 man->meta.vol = mandoc_strdup(n->string);
 389         else if ('\0' != man->meta.msec[0] &&
 390             (NULL != (p = mandoc_a2msec(man->meta.msec))))
 391                 man->meta.vol = mandoc_strdup(p);
 392
 393         /*
 394          * Remove the `TH' node after we've processed it for our
 395          * meta-data.
 396          */
 397         man_node_delete(man, man->last);
 398 }
 399
 400 static void
 401 post_nf(CHKARGS)
 402 {
 403
 404         check_eq0(man, n);
 405
 406         if (MAN_LITERAL & man->flags)
 407                 mandoc_msg(MANDOCERR_NF_SKIP, man->parse,
 408                     n->line, n->pos, "nf");
 409
 410         man->flags |= MAN_LITERAL;
 411 }
 412
 413 static void
 414 post_fi(CHKARGS)
 415 {
 416
 417         check_eq0(man, n);
 418
 419         if ( ! (MAN_LITERAL & man->flags))
 420                 mandoc_msg(MANDOCERR_FI_SKIP, man->parse,
 421                     n->line, n->pos, "fi");
 422
 423         man->flags &= ~MAN_LITERAL;
 424 }
 425
 426 static void
 427 post_UC(CHKARGS)
 428 {
 429         static const char * const bsd_versions[] = {
 430             "3rd Berkeley Distribution",
 431             "4th Berkeley Distribution",
 432             "4.2 Berkeley Distribution",
 433             "4.3 Berkeley Distribution",
 434             "4.4 Berkeley Distribution",
 435         };
 436
 437         const char      *p, *s;
 438
 439         n = n->child;
 440
 441         if (NULL == n || MAN_TEXT != n->type)
 442                 p = bsd_versions[0];
 443         else {
 444                 s = n->string;
 445                 if (0 == strcmp(s, "3"))
 446                         p = bsd_versions[0];
 447                 else if (0 == strcmp(s, "4"))
 448                         p = bsd_versions[1];
 449                 else if (0 == strcmp(s, "5"))
 450                         p = bsd_versions[2];
 451                 else if (0 == strcmp(s, "6"))
 452                         p = bsd_versions[3];
 453                 else if (0 == strcmp(s, "7"))
 454                         p = bsd_versions[4];
 455                 else
 456                         p = bsd_versions[0];
 457         }
 458
 459         free(man->meta.source);
 460         man->meta.source = mandoc_strdup(p);
 461 }
 462
 463 static void
 464 post_AT(CHKARGS)
 465 {
 466         static const char * const unix_versions[] = {
 467             "7th Edition",
 468             "System III",
 469             "System V",
 470             "System V Release 2",
 471         };
 472
 473         const char      *p, *s;
 474         struct man_node *nn;
 475
 476         n = n->child;
 477
 478         if (NULL == n || MAN_TEXT != n->type)
 479                 p = unix_versions[0];
 480         else {
 481                 s = n->string;
 482                 if (0 == strcmp(s, "3"))
 483                         p = unix_versions[0];
 484                 else if (0 == strcmp(s, "4"))
 485                         p = unix_versions[1];
 486                 else if (0 == strcmp(s, "5")) {
 487                         nn = n->next;
 488                         if (nn && MAN_TEXT == nn->type && nn->string[0])
 489                                 p = unix_versions[3];
 490                         else
 491                                 p = unix_versions[2];
 492                 } else
 493                         p = unix_versions[0];
 494         }
 495
 496         free(man->meta.source);
 497         man->meta.source = mandoc_strdup(p);
 498 }
 499
 500 static void
 501 post_vs(CHKARGS)
 502 {
 503
 504         if (n->tok == MAN_br)
 505                 check_eq0(man, n);
 506         else
 507                 check_le1(man, n);
 508
 509         if (NULL != n->prev)
 510                 return;
 511
 512         switch (n->parent->tok) {
 513         case MAN_SH:
 514                 /* FALLTHROUGH */
 515         case MAN_SS:
 516                 mandoc_vmsg(MANDOCERR_PAR_SKIP, man->parse, n->line, n->pos,
 517                     "%s after %s", man_macronames[n->tok],
 518                     man_macronames[n->parent->tok]);
 519                 /* FALLTHROUGH */
 520         case MAN_MAX:
 521                 /*
 522                  * Don't warn about this because it occurs in pod2man
 523                  * and would cause considerable (unfixable) warnage.
 524                  */
 525                 man_node_delete(man, n);
 526                 break;
 527         default:
 528                 break;
 529         }
 530 }